Building Applications in the Cloud: Best Practices
summary
High performers deploy daily or more, with lead time under an hour and mean time to recovery roughly 96x faster than low performers; aim for deploy once a day, lead time < 1 hour. Use cloud-native patterns: stateless services, no root access, decoupled APIs, immutable containers, version-control infrastructure and standardized deployment/testing patterns. Practice isolation testing (cuts UAT time and speeds sign-off), continuous delivery to production-like environments, run chaos tools like Chaos Monkey or Platform.sh’s uni kitty, and consider commodification once ≈200 containers per engineer.
AI-written summary · about
chapters (15)
- 0:00Start
- 6:44Why did the chicken cross the road?
- 7:25Why was the project late?
- 7:42Plan Do Check Adjust
- 10:14High vs low
- 18:07Optimising batch size
- 20:03A short story about testing
- 22:09Isolation testing leads to...
- 24:52Antifragile systems
- 25:21Mean Time Between Failures
- 25:35Mean Time To Recovery
- 25:58Chaos Animals
- 27:20Production-like environments
- 27:42The Continuous Deployment Stack
- 29:00Commodification
related talks
36mPractical Cloud Native: What Works, What Doesn't
45mBuilding and Scaling High Performing Technology Organizations
23mHow to Build and Deploy Services in the Cloud fireside chat
46mDeploying and maintaining software for today's infrastructure
42mSpeeding Up Innovation
36mIs DevOps Still Relevant in a Cloud Native World?
30mThe transformative power of modern applications
35mAre We There Yet? Signposts On Your Journey To Awesome